An interesting property of Searle's argument is that, if true, it means that the "strong AI" program is worthless; no discrete system that involves following rules (including neural networks and statistical models) can really "understand" anything.
The reason this came to mind is that Searle includes a large section addressing most (possibly all) of the responses to his argument by counter-counter-arguments. The SEP (https://plato.stanford.edu/entries/chinese-room/) has a lovely discussion of the problem and responses since 1980, which seems to fulfil your desire for a "concise summary of what exactly is the current state of art answer". (Unfortunately, that isn't the case in general; the SEP page on Zhuangzi (https://plato.stanford.edu/entries/zhuangzi/) simply presents one approach, although it's one I haven't seen anywhere else.)
Unfortunately, while Searle presents counter-refutations to the refutations of his argument, and others since have added more and deeper levels of refutations, none have, to my knowledge, been universally acknowledged as prevailing. I personally don't buy Searle's argument (if you accept it, and especially his response to the "systems response", you will have to admit that no one understands Chinese), but that's neither here nor there; others do, and others find failures in his argument that I think are silly.
This is the case with most all philosophy. The only way for a philosophical point to be incorrect is for it to be logically inconsistent with other parts of the same argument, and that is almost never the case in any real arena. A similar sort of situation holds occasionally over in mathematics, say, where Euclid's fifth axiom was argued over for most of 2000 years; eventually it was discovered that you can invert it and produce a similar system that is intuitively inobvious, but logically consistent (and useful). Likewise, which philosophical points you accept depends entirely on what axioms you start with (and unfortunately, philosophical axioms are harder to pin down than geometry's).
My bottom line is that the historical discussion is the point of philosophy. Understanding complex arguments, even if they aren't "true" or if you don't agree with them, is considerably more than worthless.
An interesting property of Searle's argument is that, if true, it means that the "strong AI" program is worthless; no discrete system that involves following rules (including neural networks and statistical models) can really "understand" anything.
The reason this came to mind is that Searle includes a large section addressing most (possibly all) of the responses to his argument by counter-counter-arguments. The SEP (https://plato.stanford.edu/entries/chinese-room/) has a lovely discussion of the problem and responses since 1980, which seems to fulfil your desire for a "concise summary of what exactly is the current state of art answer". (Unfortunately, that isn't the case in general; the SEP page on Zhuangzi (https://plato.stanford.edu/entries/zhuangzi/) simply presents one approach, although it's one I haven't seen anywhere else.)
Unfortunately, while Searle presents counter-refutations to the refutations of his argument, and others since have added more and deeper levels of refutations, none have, to my knowledge, been universally acknowledged as prevailing. I personally don't buy Searle's argument (if you accept it, and especially his response to the "systems response", you will have to admit that no one understands Chinese), but that's neither here nor there; others do, and others find failures in his argument that I think are silly.
This is the case with most all philosophy. The only way for a philosophical point to be incorrect is for it to be logically inconsistent with other parts of the same argument, and that is almost never the case in any real arena. A similar sort of situation holds occasionally over in mathematics, say, where Euclid's fifth axiom was argued over for most of 2000 years; eventually it was discovered that you can invert it and produce a similar system that is intuitively inobvious, but logically consistent (and useful). Likewise, which philosophical points you accept depends entirely on what axioms you start with (and unfortunately, philosophical axioms are harder to pin down than geometry's).
My bottom line is that the historical discussion is the point of philosophy. Understanding complex arguments, even if they aren't "true" or if you don't agree with them, is considerably more than worthless.